Open Access Open Access  Restricted Access Subscription or Fee Access

Combination of Feature Sets Based on Binary Pattern and Oriented Gradient for Efficient Face Classification

Bouchra Nassih, Aouatif Amine, Mohammed Ngadi, Nabil Hmina

Abstract



Generally, the face is one of the most important research, especially, face classification in pattern recognition. In this paper, based on the Local Binary Pattern (LBP) descriptor and its Uniform Local Binary Pattern (ULBP) variant, we propose to merge a set of different descriptors features which consist of Histograms of Oriented Gradient (HOG), Global Discrete Cosine Transform (GDCT), and Local Discrete Cosine Transform (LDCT). First, we
used ULBP, SLBP, and each previously mentioned techniques separately. Then, we combined them with LBP algorithm with two versions ULBP and SLBP. Afterward, we exploit them as an input of varied classification methods, as Polynomial-Support Vector Machine (Polynomial-SVM), RBF-SVM, Linear-SVM, K Nearest Neighbors (KNN) and AdaBoost. We evaluate the performance of our approach using two facial databases. The first one is BOSS. The second one is MIT-CBCL database. The experimental results reveal that the fusion features of LBP and HOG descriptors with SVM classifier exhibits robust performance
in terms of accuracy and computation time in both databases, even if, the HOG achieved a good classification result when we used it alone with 99;9% on BOSS database, which demonstrates the robustness of our database.

Keywords


Face Classification, Local Binary Pattern, Histograms of Oriented Gradient

Full Text:

PDF


Disclaimer/Regarding indexing issue:

We have provided the online access of all issues and papers to the indexing agencies (as given on journal web site). It’s depend on indexing agencies when, how and what manner they can index or not. Hence, we like to inform that on the basis of earlier indexing, we can’t predict the today or future indexing policy of third party (i.e. indexing agencies) as they have right to discontinue any journal at any time without prior information to the journal. So, please neither sends any question nor expects any answer from us on the behalf of third party i.e. indexing agencies.Hence, we will not issue any certificate or letter for indexing issue. Our role is just to provide the online access to them. So we do properly this and one can visit indexing agencies website to get the authentic information.